base class for models of PixelLocatorConceptPixel locator is similar to a pixel iterator, but allows for 2D navigation of pixels within an image view. More...

#include <locator.hpp>

Inheritance diagram for boost::gil::pixel_2d_locator_base< Loc, XIterator, YIterator >:

[legend]

Classes
struct	axis

Public Types
typedef XIterator	x_iterator

typedef YIterator	y_iterator

typedef std::iterator_traits < x_iterator >::value_type	value_type

typedef std::iterator_traits < x_iterator >::reference	reference

typedef std::iterator_traits < x_iterator > ::difference_type	coord_t

typedef point2< coord_t >	difference_type

typedef difference_type	point_t

typedef point_t::template axis < 0 >::coord_t	x_coord_t

typedef point_t::template axis < 1 >::coord_t	y_coord_t

typedef difference_type	cached_location_t

Public Member Functions
bool	operator!= (const Loc &p) const

x_iterator	x_at (x_coord_t dx, y_coord_t dy) const

x_iterator	x_at (const difference_type &d) const

y_iterator	y_at (x_coord_t dx, y_coord_t dy) const

y_iterator	y_at (const difference_type &d) const

Loc	xy_at (x_coord_t dx, y_coord_t dy) const

Loc	xy_at (const difference_type &d) const

template<std::size_t D>
axis< D >::iterator &	axis_iterator ()

template<std::size_t D>
axis< D >::iterator const &	axis_iterator () const

template<std::size_t D>
axis< D >::iterator	axis_iterator (const point_t &p) const

reference	operator() (x_coord_t dx, y_coord_t dy) const

reference	operator[] (const difference_type &d) const

reference	operator* () const

Loc &	operator+= (const difference_type &d)

Loc &	operator-= (const difference_type &d)

Loc	operator+ (const difference_type &d) const

Loc	operator- (const difference_type &d) const

cached_location_t	cache_location (const difference_type &d) const

cached_location_t	cache_location (x_coord_t dx, y_coord_t dy) const

Static Public Attributes
static const std::size_t	num_dimensions =2

Friends
template<typename X >
class	pixel_2d_locator

Detailed Description

template<typename Loc, typename XIterator, typename YIterator>
class boost::gil::pixel_2d_locator_base< Loc, XIterator, YIterator >

base class for models of PixelLocatorConcept

Pixel locator is similar to a pixel iterator, but allows for 2D navigation of pixels within an image view.

It has a 2D difference_type and supports random access operations like:

difference_type offset2(2,3);
locator+=offset2;
locator[offset2]=my_pixel;

In addition, each coordinate acts as a random-access iterator that can be modified separately: "++locator.x()" or "locator.y()+=10" thereby moving the locator horizontally or vertically.

It is called a locator because it doesn't implement the complete interface of a random access iterator. For example, increment and decrement operations don't make sense (no way to specify dimension). Also 2D difference between two locators cannot be computed without knowledge of the X position within the image.

This base class provides most of the methods and typedefs needed to create a model of a locator. GIL provides two locator models as subclasses of pixel_2d_locator_base. A memory-based locator, memory_based_2d_locator and a virtual locator, virtual_2d_locator. The minimum functionality a subclass must provide is this:

class my_locator : public pixel_2d_locator_base<my_locator, ..., ...> {  // supply the types for x-iterator and y-iterator
       typedef ... const_t;                      // read-only locator
       template <typename Deref> struct add_deref {
           typedef ... type;                     // locator that invokes the Deref dereference object upon pixel access
           static type make(const my_locator& loc, const Deref& d);
       };
       my_locator();
       my_locator(const my_locator& pl);
       // constructors with dynamic step in y (and x). Only valid for locators with dynamic steps
       my_locator(const my_locator& loc, coord_t y_step);
       my_locator(const my_locator& loc, coord_t x_step, coord_t y_step, bool transpose);
       bool              operator==(const my_locator& p) const;
       // return _references_ to horizontal/vertical iterators. Advancing them moves this locator
       x_iterator&       x();
       y_iterator&       y();
       x_iterator const& x() const;
       y_iterator const& y() const;
       // return the vertical distance to another locator. Some models need the horizontal distance to compute it
       y_coord_t         y_distance_to(const my_locator& loc2, x_coord_t xDiff) const;
       // return true iff incrementing an x-iterator located at the last column will position it at the first 
       // column of the next row. Some models need the image width to determine that.
       bool              is_1d_traversable(x_coord_t width) const;
};

Models may choose to override some of the functions in the base class with more efficient versions.